AI looks increasingly useless in telecom and anywhere else

A recent article challenges the prevailing narrative surrounding generative AI (GenAI), asserting that overreliance on these tools can lead to cognitive decline and that the highly anticipated GPT-5 is merely an incremental improvement, failing to deliver an AGI-like breakthrough. Critically, a new MIT report indicates that 95% of organizations investing $30-40 billion in GenAI are currently seeing "zero return," contributing to recent tech stock declines. The piece also debunks claims of widespread AI-driven job displacement, noting that major tech firms have significantly expanded workforces, with telecom and tech sector layoffs primarily stemming from other factors like sub-sector downturns or efficiency measures, while telcos' AI adoption has yet to yield tangible revenue growth or profitability.

Analysis

The prevailing narrative of generative AI's immediate, transformative impact is being challenged by mounting evidence of its current limitations and questionable return on investment. The recent release of GPT-5, described as a mere incremental improvement rather than a leap towards Artificial General Intelligence (AGI), underscores this sentiment, with the model failing basic causality tests. This technological disappointment is compounded by a stark financial reality highlighted in a new MIT report: 95% of organizations have achieved zero return on a collective $30-40 billion investment in GenAI, triggering a downturn in related tech stocks. Furthermore, the article debunks the notion of widespread AI-driven layoffs, revealing that tech giants like Microsoft and Alphabet have substantially increased their workforces since 2018, with job cuts at firms like Nokia and Intel attributable to market-specific downturns and company crises rather than AI implementation. A clear bifurcation is emerging where infrastructure providers—vendors of data center connectivity like Arista and Ciena, and chip designers like AMD and Nvidia—are benefiting from the traffic generated by LLM training, while enterprise adopters, particularly in the telecom sector, have yet to see any tangible boost to revenue or profitability.

Key Decisions for Investors

• Investors should critically scrutinize the valuation of companies promising near-term revenue from enterprise GenAI software, as the reported 95% rate of zero return on investment suggests a significant disconnect between market hype and current enterprise value.
• Consider overweighting positions in the "picks-and-shovels" of the AI build-out, such as data center hardware suppliers and specialized chip designers like Nvidia and AMD, which benefit from infrastructure demand regardless of the ultimate profitability of AI applications.
• Monitor telecom sector margins closely, as their AI investments are currently driving up data traffic and network costs without a corresponding increase in revenue, potentially signaling future profitability pressures.
• Exercise caution with companies like Nokia and Intel, as the analysis indicates their workforce reductions stem from fundamental business challenges and market share loss, not a successful AI-driven transformation that would imply future efficiency gains.