Big tech capex is set to jump over 50% in 2024, fueling the current AI boom, and supporting the training and deployment of the next-generation of frontier models slated for release over the next 2-4 months

If these frontier models can deliver greater capabilities, and the returns to match, it will intensify the race to scale up capex even further to train ever more powerful models on ever larger clusters of chips

If returns do not flow to the frontier, then models become commoditised, with all of big tech able to capitalise on their application layer dominance. If they do, then outcomes are uneven and uncertain with the core cloud players racing for dominance and leaving the others behind

Sony PlayStation 5 and Microsoft Xbox unit sales crashed in the last quarter, despite promotional discounting. Neither company appears able to reverse a clear consumer shift away from fixed consoles. 

Nintendo Switch outsold Xbox and nearly matched PlayStation sales, even with the Switch 2 set to launch in 2025. 

A radical change in hardware strategy and leadership will be the best solution for Microsoft to demonstrate a growth narrative following its acquisition of Activision. Gamescom this month may provide more clues.

We forecast broadcaster viewing share to drop to 52% in 2030 (from 58% in 2023), with the firming of its on demand viewing unable to balance out the decline of live: this is a slight improvement on our past estimates, with decline slowing.

SVOD viewing will begin to plateau in 2025, as video sharing platforms (YouTube, TikTok, Twitch) take an increasing share of engagement.

On the TV set, YouTube will grow strongly: we predict a 90% increase from 2023 to 2030. This is from a low base with broadcasters retaining 70% of viewing on the main screen in 2030

Meta led the pack of tech results in Q2 with 22% growth and championing a suite of generative AI products; should these falter, Meta can recalibrate by devoting more of its AI infrastructure to core user and ad products. 

AI and the metaverse give Meta an uncertain shot at a new platform play, leveraging its enormous user base and bringing developers back into the fold. 

Reality Labs is still burning cash, but a collaboration with Ray-Ban offers a path to usable head-mounted displays, and could get Meta there faster than Apple’s cutting-edge approach. 

The next generation of the largest and most powerful 'frontier' AI models will be a key test for the pace of AI progress, with OpenAI's upcoming GPT-5 the most highly anticipated.

For OpenAI, the stakes are high, facing a growing assortment of rivals and with huge spend on training and running models to recoup. Staying at the cutting edge is key to justifying itself to the big tech backers on which it depends.

If OpenAI can deliver technology that matches its ambitious vision for what AI can be, it will be transformative for its own prospects, but also the economy more broadly. Falling short could be fatal.

AI integration into production tools throughout media industries will deliver increased productivity for professional content creation. Generally available tools will also improve quality and production speed for individual user-creators.

Roadblocks include the uncertain copyright status of models and their outputs, attitudes of creative workers and consumers, and the AI tech underdelivering versus what was promised. The need to integrate new tools into existing processes is perhaps the biggest brake.

There are stark differences by sector: the opportunities are greatest in games, where costs have ballooned and software engineering is core. Marketing is furthest in exploiting AI, while audiovisual production is more cautious.

On 4 June 2024, Enders Analysis co-hosted the annual Media and Telecoms 2024 & Beyond Conference with Deloitte, sponsored by Barclays, Salesforce, the Financial Times, and Adobe. 

With over 580 attendees and over 40 speakers from the TMT sector, including leading executives and industry experts, the conference focused on how new technologies, regulation and infrastructure will impact the future of the industry. 

This is the edited transcript of Session Four, covering: artificial intelligence, the new phase of online advertising, and closing remarks. Videos of the presentations are available on the conference website.

TikTok has been dealt a devastating blow as a US bill has been signed into law forcing owner ByteDance to sell within a year or face its removal from app stores. 

The stakes are higher than in 2020—China's opposition to a divestment will make an optimal sale harder to conclude, so all sides must be prepared for a ban.   

The TikTok bill introduces extraordinary new powers in the context of the US and China's broad systemic rivalry, though online consumer benefits will be limited.  

Big news publishers are pursuing licensing deals with AI companies, chiefly OpenAI. Not all publishers will see a substantial return; while some news may be important for training AI models, not all publisher content will be

Litigation is a threat point when negotiations stall (see the New York Times), but the copyright status of Large Language Models (LLMs) is uncertain. In the UK, there has been no government intervention (on copyright or otherwise) that could facilitate licensing 

Publishers’ bargaining position is strongest when it comes to up-to-date material that could be important in powering some AI consumer products. They should seek deals to support their journalism, while bearing in mind the risk that new products may get between them and their readers

 

The US is intent on preventing the CCP’s goal of AI supremacy by 2030, banning exports of advanced AI chips to Chinese companies. So far, these bans have largely been shrugged off to create a new commercial dynamic in the region. 

Huawei wields a de facto monopoly on the manufacture and sale of advanced chips in China. Huawei also sells cloud services globally and threatens Apple's $70 billion in Chinese revenues through its premium handsets. 

China’s AI regulation is highly supportive of the training and deployment of Chinese-language LLMs developed by tech platforms, startups, and device makers, with meaningful revenue gains only appearing by H2 2024.