Headlines this week - Jun 22, 2025

A look at how capital is being deployed across future opportunities

C Gavilanes

Jun 22, 2025

This week in the future:

1 - AI must foster creativity, not just automate jobs, to solve the productivity paradox

History suggests that efficiency is not enough to catalyze productivity growth. To overcome the last 30 years of stagnating labor-productivity growth, AI must be a tool for creativity, not just efficiency. According to an analysis by Oxford professor C. B. Frey, simply replacing jobs is not enough. While technical breakthroughs are necessary, the greater challenge will be the institutional changes required to integrate these technologies into creative work
"Hybrid" jobs, where humans use AI to enhance their skills, may be part of the solution. Amazon CEO A. Jassy recently encouraged employees to "be curious" about AI, arguing that those who use it effectively are best positioned to "reinvent the company".

Meanwhile, fear of job losses could hinder the transition. Recent data shows a negative impact on employment for recent US college graduates, whose unemployment rate is 2.6 percentage points above the national average. These signals have already started to create concern

2 - Meta's investment in Scale AI might signal that talent and data are the new competitive battleground

One week after the deal, the market reaction appears positive, reflecting a belief that Meta can leverage AI to better monetize its user data

The Scale AI deal is viewed as a way for Meta to create a competitive “moat” vs. other labs. An FT opinion column this week argues that superior talent and data (which Scale AI provides) are the key differentiation levels for AI labs, given that “data centers and the chips that fill them are a commodity, albeit a very expensive one” (but I tend to disagree that an “expensive commodity” can’t be used as a competitive advantage…)

On top of this, Scale AI could also be a source of competitive intelligence. The previous commercial relationships of Scale AI with most of the other AI labs, and the information that the company has obtained from that, could give Meta leverage over its competitors, according to a Bloomberg article

Seen at an industry level, the deal could be the beginning of a round of M&A deals to acquire data specialists. In a Reuters article this week we read that what the author calls “legacy tech companies” are interested in acquiring specialists exposed to management of clients’ data in the cloud. This could be key to accelerate AI adoption in the business segment

3 - SoftBank plans a $1trn AI and robotics hub in Arizona in partnership with TSMC

The ambitious plan, codenamed "Project Crystal Land," aims to build AI data centers and robot manufacturing facilities. While SoftBank representatives are already in discussions with US government officials, TSMC's potential role in the project remains unclear.
To fund its AI ambitions, SoftBank is reducing its telecom exposure. They have recently sold a $4.8bn stake in T-Mobile US

4 - Powering AI data centers is expected to require a massive coordinated effort from industry and government

The challenge is comparable with the one for the space race in the 60s, according to the ex-CEO of BP, who now chairs Prometheus Hyperscale (an AI data center startup)

It is not at all clear that the current grid will be able to cope with the expected demand growth, specially if it is run by the same companies in the same way

Policy will be a key issue. E.g. a Wired article this week argues that these AI companies remain reluctant to reveal their actual energy needs, that are still covered, to a large extent, through fossil fuels (e.g. natural gas) for fear of the reaction of climate activists

5 - Innovation across the biotech sector is rapidly gaining momentum. Biotechnology is seeing breakthroughs on multiple fronts, creating significant investor excitement

Market rally in China. The local Hang Seng Biotech Index has surged over 60% since January, outperforming the tech sector's 17% gain (including the “DeepSeek effect”). This has been driven by major licensing deals for new Chinese drugs by foreign firms
More startups using AI for drug discovery. SandboxAQ, a Google spin-off backed by Nvidia, just released a synthetic dataset to train AI models that predict how a drug's structure will bind to proteins. The company has already raised over $1 billion
Genomics acceleration. Startups like Nucleus are leveraging computing power to build an "application layer for genomics". By testing the whole DNA sequence, they can offer richer health insights than services like 23andMe

A new wave of biotech M&A deals. As an example, we have the recent acquisition of Verve, a CRISPR / Gene-editing startup by Eli Lilly, for $1.3bn

A note of caution: fertility stocks. The FT’s Lex Column explains that listed fertility firms have performed poorly due to (1) some companies’ “noisy” failures (e.g. mixing up embryos); (2) demand overestimation (for an expensive service); (3) challenges to improve companies’ margins, given their high operational costs

6 - SpaceX's Starship rocket suffers a setback with an explosion during a test.

The company attributes the accident to “a major anomaly” while on a test stand. There were no casualties, but these are obviously not good news for the project, critical both for NASA’s next manned mission to the Moon (Artemis) and for an eventual future mission to Mars

7 - France moves to control OneWeb, aiming to build Europe’s alternative to Starlink

The French government is investing €717m in satellite operator Eutelsat, increasing its stake from 13.6% to nearly 30%. The move is designed to secure control over Eutelsat's key asset, OneWeb, which France sees as a regional satellite communications alternative to SpaceX's Starlink. As France's influence grows, the UK's stake in the company is being diluted from 11% to 7.9%

8 - The robotaxi race heats up with new launches in major US cities.

Tesla is launching its first robotaxi service in Austin, Texas, this weekend with an initial fleet of 10 cars. The company claims its reliance on cameras and software upgrades for its existing consumer vehicles gives it a better path to scale and profitability than competitors like Waymo, who use more expensive LIDAR sensors

Waymo is planning to launch in New York City. but must first address a state law that prohibits vehicles from operating without a person at the wheel. Service tests will begin next month with human safety drivers on board.

9 - Nvidia invests in a nuclear energy startup (founded by Bill Gates)

Nvidia's corporate VC fund has participated in a $650m funding round for TerraPower, a nuclear energy company founded by Bill Gates. TerraPower is currently building a 345MW reactor in Wyoming. South Korea's Hyundai is also an investor in the round

10 - Global interest is growing on brain-computer interfaces

China enters the race with its first successful clinical trial. Chinese researchers have conducted the country’s first clinical trial of a brain-computer interface (BCI) technology, allowing a patient with tetraplegia to control electronic devices with his mind