Breaking news

Google Cloud VP Questions Long-Term Viability Of LLM Wrapper

Rethinking AI Startup Business Models

The rapid growth of generative AI has produced a wave of startups, but some early business models are now facing increased scrutiny. Companies built primarily as wrappers around large language models such as Claude, GPT, or Gemini are being questioned over their limited proprietary technology.

Insights From A Cloud Veteran

Darren Mowry, Vice President of Global Startups at Google Cloud, discussed these dynamics during an episode of TechCrunch’s Equity podcast. He said startups relying on a simple interface layered on top of an existing language model may struggle to differentiate. According to Mowry, packaging third-party AI models without building proprietary capabilities becomes difficult to sustain once cloud credits expire and operating costs increase.

Beyond Wrappers: The Aggregator Dilemma

AI aggregators, which combine multiple large language models under a single interface or API, face similar pressure. While these platforms often offer orchestration tools such as monitoring, governance, and evaluation, investors and customers are increasingly focused on products with clear intellectual property. Mowry advised founders to avoid the aggregator model unless it includes meaningful technical differentiation.

Parallels With Early Cloud Innovation

Mowry compared the current AI cycle to the early cloud computing era. At that time, many companies attempted to resell AWS infrastructure but struggled once Amazon launched its own enterprise tools. Firms that survived expanded into areas such as security, migration, and DevOps services. He suggested AI startups follow a similar path by building deeper value beyond access to foundational models.

Emerging Opportunities In AI And Beyond

Despite concerns around wrappers and aggregators, Mowry pointed to strong momentum in developer platforms and direct-to-consumer tools. Companies such as Replit, Lovable, and Cursor have gained traction through product differentiation and user adoption. He also highlighted growth in sectors outside core AI, including biotech and climate tech, where data-driven innovation is generating new opportunities.

Building For Long-Term Success

The current market environment favors startups that develop defensible advantages through vertical specialization or clear product differentiation. Founders who rely solely on existing backend models may struggle to maintain long-term competitiveness.

For startups operating in a rapidly evolving AI ecosystem, sustained success depends on building proprietary value and scalable business fundamentals.

Women Make Up A Majority Of The EU’s Science And Technology Workforce But The Real Gap Is Elsewhere

Women now make up the majority of the EU’s science and technology workforce. According to Eurostat, in 2025, more than 81.6 million people aged 15 to 74 were employed in science and technology occupations across the EU. Of those, 52.5% were women, equal to 42.8 million women. The number of women in these occupations rose by 27.9% compared with 2015, an increase of more than 9.3 million over a decade.

On the surface, the numbers resemble progress. However, Eurostat’s category requires context before that figure can be read accurately. The data refers to HRST, or Human Resources in Science and Technology, specifically people employed in science and technology occupations. These are roles where the main tasks require professional or technical knowledge in physical and life sciences, but also in social sciences and humanities. That definition is wider and broader than engineering, ICT, laboratory science, or high-tech research alone.

Zooming In

The gender picture changes once the data moves from a wider definition of the workforce to the narrower scientist-and-engineer (research and manufacturing) subgroup.

Scientists and engineers represented almost a quarter of all people employed in science and technology in the EU in 2025. Eurostat describes scientists and engineers as often being the innovators at the centre of technology-led development, making them an important subgroup to focus on separately.

Women accounted for only 40.8% of scientists and engineers in 2025, despite making up more than half of the wider category. That share has increased by a mere 0.5 percentage points over the past decade. The absolute number of women working as scientists and engineers rose from 5.3 million in 2015 to 8.2 million in 2025, despite the push from national and international organisations to increase the number of women in the field. Europe has expanded the number of women in science and technology occupations over ten years. However, that expansion has not extended equally into the scientist-and-engineer subgroup, where much of Europe’s research and innovation work is conducted.

In 2025, of the 39.4 million women aged 25 to 64 working in science and technology occupations in the EU, 35.5 million worked in service activities. Only 2.7 million worked in manufacturing. Women accounted for 57.5% of science and technology employment in services, but only 31.3% in manufacturing.

In 2025, the highest shares of women employed in science and technology occupations were recorded in Latvia at 62.4%, followed by Hungary’s Great Plain and North region at 61.1%, Estonia at 60.5%, Poland’s Central macroregion at 60.4%, and Lithuania at 60.3%. No EU country recorded a majority of women among science and technology workers in manufacturing.

Break-down

Eurostat’s figures measure employment in broad science and technology occupations. They do not show job security, pay levels, management roles, promotion rates, research leadership, or whether women are concentrated in junior or senior workplace positions.

The classification of “senior” also requires additional explanation. Eurostat reports that 45.9% of science and technology workers aged 25 to 64 in the EU were classified as “senior” HRST in 2025. In this dataset, “senior” refers to workers aged 45 to 64. It does not mean senior manager, senior researcher, team lead, or decision-maker.

A high female share in the wider Human Resource Science and Technology (HRST) category does not parallel equal representation across scientists, engineers, manufacturing roles, senior posts, pay, research funding, or decision-making. These figures also reflect the occupational mix inside each country or region, not only structural progress across all areas of science and technology.

The Case Of Cyprus

Eurostat data places Cyprus’s overall science and technology employment at 37.2% of the labour force in 2025, slightly above the EU-27 figure of 36.9%, and above Greece at 26.8%, Malta at 33.9%, and Turkey at 18.2%. This figure covers the total share of the labour force employed in science and technology across all genders.

Progress Or Work-in-Progress?

52.5% in the broad category. 40.8% among scientists and engineers. 31.3% in manufacturing. Europe’s gender gap in science and technology hasn’t closed yet, and there is still work to be done to encourage and support more women to enter the field, especially in research and manufacturing.

Let’s not wait another decade for another couple of percentage points of hope.

The Future Forbes Realty Global Properties
Uol
eCredo
Aretilaw firm

Become a Speaker

Become a Speaker

Become a Partner

Subscribe for our weekly newsletter