Ghada Ismail

As artificial intelligence becomes increasingly embedded in business, not everything an AI system generates should be taken at face value.

Two concepts often create confusion in this context: synthetic data and AI hallucination. Both involve information generated by AI rather than directly collected from the real world, but their roles could not be more different.

One is a tool that can help businesses overcome data limitations. The other is a reliability problem that can undermine trust in AI systems.

What Is Synthetic Data?

Synthetic data is artificially generated information designed to replicate the characteristics and patterns of real-world data.

Instead of collecting thousands of real customer transactions, for example, a startup could generate synthetic transactions that mimic realistic purchasing behavior. Similarly, an AI developer could create synthetic images, customer profiles or financial scenarios to train and test an AI model.

This can be particularly valuable for startups that lack access to large datasets or operate in areas where data is sensitive.

Synthetic data can help companies reduce data-collection costs, accelerate AI development and limit exposure to sensitive information. It can also allow developers to test AI systems across scenarios that may be difficult or expensive to reproduce in the real world.

However, synthetic data is only useful when it is representative and properly validated. Poor-quality synthetic datasets can reproduce errors, biases or unrealistic patterns.

What Is AI Hallucination?

AI hallucination is something very different.

It occurs when an AI model generates information that sounds convincing but is factually incorrect, unsupported, or completely fabricated.

An AI chatbot, for instance, might invent a statistic, cite a research paper that does not exist, or provide an incorrect explanation with complete confidence.

Hallucinations can occur because generative AI models are designed to predict and generate likely sequences of information. They do not automatically distinguish between what is true and what merely appears plausible.

For businesses, this can become a serious issue. An inaccurate AI-generated answer may be inconvenient in a consumer application but potentially damaging in areas such as financial services, healthcare, legal technology or enterprise decision-making.

Synthetic Data vs AI Hallucination

The simplest way to distinguish the two is intention and purpose.

Synthetic data is deliberately created. AI hallucination is an unintended output.

Synthetic data is generated for a specific purpose, such as training, testing, or simulating scenarios. It can be reviewed, measured, and validated before being used.

Hallucinations, by contrast, emerge during an AI system's operation and need to be detected, corrected, or prevented.

In other words, synthetic data can be an AI development asset, while hallucination is an AI reliability risk.

Why Does This Matter for Startups?

The distinction is especially important for startups building AI products.

Early-stage companies often face limited access to high-quality data. Synthetic data can provide a way to experiment and develop models without relying exclusively on costly or sensitive real-world datasets.

At the same time, startups must ensure that their AI products do not generate unreliable information. A hallucination can quickly erode customer confidence, particularly when an AI product is being used to make business or financial decisions.

Importantly, synthetic data does not automatically cause hallucinations. However, if synthetic datasets are poorly designed or contain unrealistic patterns, they can affect the quality of the models trained on them.

That makes data validation, testing, and human oversight critical throughout the AI development process.

One Is a Tool, the Other Is a Risk

Synthetic data and AI hallucination may both involve AI-generated information, but treating them as interchangeable misses a crucial distinction.

Synthetic data can help startups solve one of AI's biggest challenges: access to useful, scalable, and privacy-conscious data.

Hallucinations represent another challenge: ensuring that AI systems remain accurate and trustworthy.

As businesses move beyond experimenting with AI and begin deploying it in real-world operations, knowing the difference between data that was intentionally generated and information that was unintentionally invented will become increasingly important.

Noha Gad

In the investment management world, it is common for fund managers to earn a performance fee when they generate strong profits for their clients, but this arrangement can create an unfair situation if those gains are later lost and then partially recovered. Without additional safeguards, a manager could collect a performance fee during a good year, see the portfolio value drop sharply in the following year, and then earn another performance fee simply by bringing the fund back to its earlier level even though investors have not truly benefited from any new gains.

The high-water mark is a widely used rule in hedge funds and other managed investment products that prevents this outcome by linking performance fees to real, additional value creation rather than temporary swings in portfolio value. This rule sets the highest value that the fund has ever reached as a benchmark, allowing managers to charge a performance fee only on profits that rise above that previous peak.

What is meant by a high-water mark?

This term refers to the highest level that a body of water reaches, but metaphorically, it refers to the peak value of an investment fund or the highest point of achievement.

In the business realm, the high-water mark is a benchmark investment funds use to ensure investors only pay performance fees when a fund’s value reaches a new peak. It ensures that investors do not have to pay performance fees for poor performance, but, more importantly, guarantees that investors do not pay performance-based fees twice for the same amount of performance.

For asset management companies, including a high-water mark in their fee structure can be a strong signal of fairness and alignment with investors, ultimately contributing to attracting and retaining capital in a competitive market.

From a managerial perspective, the high-water mark encourages a focus on sustainable, long-term performance rather than short-term increases that might be followed by sharp declines. As performance fees are only available after the fund exceeds its highest historical value, managers have a clear incentive to avoid strategies that generate volatile returns with large drawdowns.

Why do high-water marks matter?

High-water marks are widely viewed as a key investor protection in hedge funds and other performance-fee-based investment structures, and they bring several clear advantages for both investors and fund managers. This includes:

• Protecting investors from paying twice for the same gains.
• Aligning manager incentives with genuine outperformance.
• Promoting more disciplined risk management.
• Supporting long-term thinking over short-term spikes.
• Enhancing trust and credibility with investors.
• Encouraging clearer communication about performance.

In conclusion, the high-water mark is more than a technical fee detail; it is a core element of fair and transparent performance-based compensation in investment management. Setting the fund’s highest historical value as the threshold for performance fees ensures that managers are rewarded only for creating new gains, not for recovering past losses or simply returning to earlier levels.

For investors, this structure provides a clear safeguard against paying twice for the same performance and helps align the manager’s interests with their own long-term outcomes. For managers and firms, it encourages more disciplined risk-taking, supports a focus on sustainable growth, and can strengthen trust and credibility in a competitive market.

Ghada Ismail

Every entrepreneur starts with an idea, but an economy becomes truly entrepreneurial when those ideas translate into businesses, jobs, investment, and new industries.

For Saudi Arabia, that transition is becoming increasingly visible.

As the Kingdom marks World Entrepreneurs Day on 21 August 2026, entrepreneurship is no longer a marginal part of its economic diversification agenda. It has become one of the key mechanisms through which Saudi Arabia is seeking to build a more dynamic private sector, create employment opportunities and develop new sources of non-oil growth.

The latest figures suggest that this transformation is gathering momentum.

According to the Global Entrepreneurship Monitor (GEM), Saudi Arabia’s Total Early-stage Entrepreneurial Activity (TEA), which measures the proportion of people aged 18 to 64 who are either starting a business or running a new one, reached 28.9% in 2025, up from 26% in 2024. The rate has more than doubled from 12.1% in 2018, highlighting the rapid expansion of early-stage entrepreneurial activity over the past seven years.

That growth is supported by an even larger pool of potential entrepreneurs. Entrepreneurial intentions reached 48.5% in 2025, meaning nearly one in two working-age adults not already involved in entrepreneurial activity intended to start a business within the next three years.

The figures point to something broader than a startup boom: a shift in attitudes toward entrepreneurship itself.

GEM found that around nine in 10 adults in Saudi Arabia either know someone who has recently started a business, believe they have the skills and experience to do so, or see good opportunities to establish a company locally. The findings suggest that entrepreneurship is increasingly viewed not simply as an alternative to employment, but as a viable career and wealth-building path.

From intention to business creation

Intentions, however, only matter when they translate into businesses.

Here, Saudi Arabia's latest company formation figures provide another indication of momentum.

During the first half of 2026, 46,900 new companies were established in the Kingdom, according to the Saudi Competitiveness and Business Center. During the same six-month period, the center delivered more than 2.9 million services to businesses, registered 86,800 establishments and verified 3,500 online stores.

The numbers reflect an increasingly streamlined environment for entrepreneurs. The center now connects businesses to around 4,800 services through integration with 80 government entities, covering areas ranging from company formation and licensing to tax, zakat and commercial registration.

This infrastructure matters because entrepreneurship is shaped not only by access to capital, but also by how easy it is to turn an idea into a legally operating business.

Saudi Arabia's broader competitiveness indicators also point in the same direction. The Kingdom ranked 13th globally and third among G20 economies in the 2026 World Competitiveness Yearbook, while authorities say around 1,000 legislative, procedural and technological reforms have been implemented to improve the business environment.

Capital follows opportunity

The evolution of entrepreneurship can also be measured by the willingness of investors to back Saudi founders.

Saudi Arabia recorded its strongest venture capital year on record in 2025, with both funding and transaction activity reaching new highs, according to MAGNiTT. The Kingdom raised $1.72 billion across 257 venture capital deals, making it the largest venture capital market in MENA by both funding and deal activity.

The momentum continued into 2026, although the market became more selective.

MAGNiTT's H1 2026 Saudi Arabia Venture Capital Report found that funding declined 74% year on year to $219 million, while deal count fell 41% to 72 transactions. Despite the slowdown, Saudi Arabia remained one of MENA's most active venture markets, although its share of regional funding fell sharply from 49% in H1 2025 to 16% in H1 2026.

The changing funding landscape is important. A mature ecosystem is not necessarily one where funding rises every year. It is one where investors increasingly distinguish between scalable businesses, sustainable business models and companies that can generate long-term value.

The next challenge: building companies that last

Saudi Arabia's entrepreneurial story, therefore, is no longer simply about how many companies are being created.

The more important question is how many can survive, scale, and become major employers or regional businesses.

This is particularly relevant because GEM found that while the percentage of adults starting or running new businesses reached 28.9% in 2025, established business ownership fell to around one in eight adults, compared with around one in five a year earlier.

The gap highlights the next stage of Saudi Arabia's entrepreneurial journey: turning a high volume of early-stage activity into businesses that survive, scale and contribute to long-term economic growth.

Creating a company is only the first milestone. Entrepreneurs need access to follow-on funding, skilled talent, customers, technology and international markets if startups are to progress from early-stage ventures into durable businesses.

There are encouraging signs. Four in five Saudi new entrepreneurs surveyed by GEM anticipated employing more than five additional people within five years, pointing to strong growth and employment ambitions among the country's emerging business owners. At the same time, digital technology is becoming increasingly central to how these entrepreneurs reach customers and grow, with a similar proportion expecting to use more digital technology to sell their products in the following six months.

For World Entrepreneurs Day 2026, this may be the most important story behind the numbers.

Saudi Arabia is not simply producing more entrepreneurs. It is building the infrastructure, capital markets and institutional environment around them.

The Kingdom's next entrepreneurial chapter will be measured not only by the number of startups founded, but by the number that scale from local ideas into national champions, regional platforms and global companies.

That is where the real economic impact of Saudi entrepreneurship will ultimately be decided.

By: Ahmed Rashad, Sr. AI Specialist, Middle East & Africa at Nutanix

Have you seen this film before? A new technology arrives, powerful and effortlessly accessible. Departments spin up projects with minimal oversight from IT or finance. The first efforts reproduce old ways of working, and then somebody rethinks the workflow entirely, and the pace picks up. Then the invoice arrives, and the organization discovers it must bring things under control without cutting off access, because access is now how the work gets done.

That was the cloud, twenty years ago. It is gen AI today, on fast forward. What took cloud most of a decade is taking enterprises about eighteen months.

We watch this from two seats. We run our own AI workloads on our own GPUs, so we have made these mistakes with our own money. We also sit alongside a great many organizations making them at the same time, in different industries and under different regulatory regimes. The striking thing is how little the story varies.

Everyone’s first question is the wrong one

It is almost always “which model?”, and it is the question that matters least, because the answer changes every quarter.

The question that survives contact with production is what a unit of work costs. Not cost per token, but cost per resolved support ticket, per merged pull request, per document retrieved. The unit price keeps falling while total spend keeps climbing, because cheaper inference simply means more inference. Jevons would have recognized it immediately.

The same discipline applies to the benefit side. Where organizations measure carefully, the gains tend to land in a recognizable range: on the order of 10 to 15 percent for support teams, and 20 to 25 percent in feature delivery velocity for engineering teams. Those numbers are only worth quoting when they have been instrumented beforehand, against a baseline captured before deployment. Worth knowing: a randomized trial by METR found that experienced developers completed real tasks 19 percent slower with AI tools, while believing they had been 20 percent faster. If you cannot say how you measured, you have a feeling rather than a result.

Agents are not chatbots, and they do not fail like chatbots

This is the shift most organizations are unprepared for. A person using an assistant makes a request and receives an answer, and both the cost and the blast radius are bounded by their attention. An agent decides for itself how many steps to take, which systems to touch, and what to do with whatever it finds. The same instruction on a different day produces a different number of tool calls, a different bill, and a different set of side effects.

Which means the controls that work are the ones you would apply to a new joiner with production access, not the ones you would apply to software licenses. An identity for every agent, distinct from the human who launched it. Permissions scoped to each tool and each system, because MCP support is table stakes now, but speaking MCP and letting you grant

an agent read access there and write access nowhere are very different things. Budget ceilings that are enforced rather than alerted on. Traces detailed enough to reconstruct why an agent took eleven steps rather than three. And a human gate on anything irreversible.

The organizations getting this right have arrived at the same architectural conclusion independently. Those decisions cannot live inside each application. They belong at a single point that every agent’s requests pass through, so that policy, spend and audit are answered once for the whole estate rather than reimplemented project by project.

Running inference in production is a different discipline from running a pilot

A demo needs one model to work once. Production needs many models to work continuously, at predictable cost, while the field moves underneath you. Every organization we work with has replaced a model in production faster than it expected to, whether because of a cheaper open weight release, a regulatory constraint, or a change in vendor pricing. The ones who suffered were those who had welded a specific model to a specific location and a specific set of applications.

Flexibility here is not a luxury; it is the whole game: serving different models for different tasks, sizing endpoints to demand, and sharing GPUs across workloads through partitioning and scheduling rather than dedicating them. And, unfashionably, batch. Document classification, index rebuilds and evaluation runs do not care whether they complete at 14:00 or at 04:00. Defer them, and interactive workloads get the daytime capacity they need. Banks ran on this logic throughout the mainframe era. It was never wrong. It merely stopped being necessary when compute was cheap.

Location is becoming a variable, not a decision

Public cloud wins on speed and on access to the newest hardware. Other forces push the opposite way. Data residency and sovereignty requirements are no longer a compliance checkbox to be satisfied at the end of a project. For a growing number of organizations, they determine which workloads can exist at all, and where. Add data gravity, latency to customers, and the economics of sustained utilization, and owned or collocated infrastructure starts to look like the sensible home for a meaningful share of inference.

Meanwhile, a new class of specialized GPU providers has appeared, and some of the organizations we work with are becoming those providers themselves, turning regional advantage and spare capacity into a business of their own.

Nobody gets this allocation right at the first attempt. What matters is that getting it wrong stays cheap to correct: that a workload can move between owned, rented and regional infrastructure without being rewritten, and that governance follows it when it moves.

Do not build a walled garden

The temptation is to stand AI up as a separate estate, with its own tooling, its own rules and its own team, deliberately quarantined from everything else. There are two problems with that.

The first is that agents produce nothing of value until they can reach the systems and the data where your business actually runs. A wall built for safety very often becomes the reason a promising pilot never becomes production. The capability works. It simply is not allowed near anything that matters.

The second is the arithmetic of running everything twice. Two sets of policies, two audit trails, two places to look during an incident, and two opportunities for them to contradict each other, while the people who understand your controls best sit on the far side of the wall from the workloads that need them most.

The organizations moving fastest treat AI as a workload like any other, subject to the same access model, the same operational discipline and the same teams, with the controls that are specific to AI layered on top rather than rebuilt alongside.

Where that leaves us

There is no magic bullet for a technology moving this fast, and anyone selling one is selling something else. But the discipline transfers even when the tools do not. Measure cost per unit of work. Instrument your claims before you repeat them. Give agents identities, budgets and boundaries, enforced in one place. Keep models and workloads free to move. And govern all of it with your estate rather than beside it.

The film is on fast forward, and none of us gets to slow it down. But you can learn the genre well enough to see the twists coming, and avoid being the character who loses the plot.

Ghada Ismail

Starting a company usually means dealing with uncertainty from day one. There is no guaranteed market, no perfect product, and often no clear answer to what comes next. This is exactly where an Entrepreneur-in-Residence (EIR) can make a difference.

An EIR is an experienced entrepreneur who temporarily joins an organization such as a venture capital firm, accelerator, incubator, university, or large company. The idea is fairly simple: bring someone with real experience of building businesses into an environment where new ideas are being explored.

But an EIR is not just another adviser sitting in meetings and giving founders advice. Depending on the organization, they may be expected to find a business opportunity, test an idea, work with startups, build a product, or even create a new company.

So, What Does an EIR Actually Do?

There is no single job description for an Entrepreneur-in-Residence. The role can look very different from one organization to another.

At a venture capital firm, an EIR might spend time looking at new markets and technologies, meeting founders, helping portfolio companies, or developing a startup idea that the firm believes could have potential.

In other cases, the EIR may already have an idea. The organization provides access to its network, resources, funding, or expertise while the entrepreneur works on turning that idea into something viable.

EIR vs. Consultant: What’s the Difference?

The two roles can sound similar, but there is an important distinction. A consultant is usually brought in to solve a specific problem. They analyze the situation, provide recommendations, and move on to the next project. An EIR is generally much closer to the building process. They might spot an opportunity, test whether customers actually want the product, find potential co-founders, develop an early version of the business, and eventually launch it.

In other words, a consultant is often paid to advise, while an EIR may be expected to build.

Why Are Venture Capital Firms Interested in EIRs?

For VC firms, an EIR can be a way to create opportunities rather than simply wait for founders to walk through the door.

Experienced entrepreneurs often know how to recognize problems worth solving. They also understand what it takes to turn an early idea into a company. By bringing these people into the firm, investors can explore new sectors and business models from the inside.

There is another advantage: relationships.

An experienced entrepreneur usually brings a network of founders, engineers, executives, investors, and industry specialists. That network can be valuable when an idea starts moving from the whiteboard to the real world.

What Makes a Good EIR?

Being a successful founder is helpful, but it is not enough.

A good EIR needs to be comfortable with uncertainty. They need to know how to ask the right questions, test assumptions quickly, and recognize when an idea is not working.

Curiosity is just as important as experience. Markets change, technologies evolve, and what worked for a previous startup may not work for the next one.

Most importantly, an EIR needs to be willing to get their hands dirty. Building a company involves far more than having a good idea. It means speaking to customers, testing products, recruiting people, changing direction, and sometimes starting over.

To Wrap Things Up…

An Entrepreneur-in-Residence is essentially an experienced builder given the time, space, and resources to explore what could come next. For investors and organizations, it can be a way to uncover new opportunities while bringing entrepreneurial experience closer to the decision-making process. For entrepreneurs, it offers a chance to explore their next move without having to start entirely from zero.

As startup ecosystems become more sophisticated, the EIR model offers an interesting middle ground between building, investing, and exploring.