Growth is holding, inflation is moderating, and the labour market remains resilient; yet the composition of the data continues to matter more than the headline figure.
January CPI rose 0.2% month-on-month and 2.4% year-on-year, down from 2.7% previously. Core CPI slowed to 2.5% year-on-year. Energy, particularly gasoline, was a key contributor to the softer headline inflation, while overall goods inflation remained contained. However, service categories persist, keeping core measures elevated relative to the target. This distinction is critical because the Federal Reserve focuses primarily on PCE, not CPI, and inflation in services within PCE will determine the durability of the disinflation trend.
January releases are also seasonally sensitive. Updated seasonal factors were incorporated into this report, reinforcing the need to avoid over-interpreting a single print. Disinflation progress is visible, but it is gradual and increasingly services-driven rather than broad-based across goods.
There are early indications that tariff effects may be emerging in selected durable-goods categories, such as appliances, furniture, and electronics. The pass-through appears narrow and uneven rather than systemic. For now, energy softness continues to offset some of that pressure.
The labour market continues to show resilience. Nonfarm payrolls increased by 130,000 in January and the unemployment rate stands at 4.3%. Average hourly earnings remain in the mid 3% year-on-year range. Sectoral gains were concentrated in health care, social assistance and construction. While the headline suggests stability, prior benchmark revisions materially reduced previously reported job growth in 2025, implying that underlying momentum was weaker than earlier estimates suggested. The current trajectory appears consistent with gradual normalisation rather than renewed acceleration.
Growth data outside the United States remain modest but positive. The United Kingdom expanded by 0.1% quarter-on-quarter in Q4, reflecting limited momentum and weak investment trends. The Eurozone grew by 0.3% quarter-on-quarter, supported by steady employment growth. The global backdrop remains characterised by slow expansion rather than contraction.
Beneath aggregate strength, dispersion persists. Consumer dynamics continue to exhibit an uneven distribution, with balance-sheet exposure influencing sentiment and spending resilience. The economy appears solid at the headline level, yet internal divergences remain visible.
From a policy perspective, the softer CPI print encouraged a more dovish market interpretation. However, inflation remains above the Federal Reserve’s 2% objective and core services continue to run firm. Fiscal support remains in the background, limiting slack. The current data mix supports patience rather than an urgent easing cycle. Confirmation through sustained moderation in PCE services inflation will be essential.
The U.S. dollar has continued to weaken relative to its multi-year rally. The move appears consistent with incremental global diversification and narrowing rate differentials rather than structural capital flight. Allocation shifts at the margin are evident, but reserve dominance remains intact.
Finally, concerns regarding statistical capacity are operational rather than political. Budget and staffing constraints may increase survey volatility and revision risk over time. In a policy environment sensitive to small changes in inflation and employment trends, data quality matters.
Rates
Treasury markets experienced a sharp intra-week repricing as investors digested conflicting signals from growth and inflation data. The sequence mattered.
The week began with a stronger-than-expected employment report, which surprised investors and temporarily pushed yields higher. The surprise resilience in the labour market led traders to push back expectations for the first rate cut from June to July, sending the 2Y yield up as much as 9.5 basis points to 3.55%, before easing back toward 3.50%. Rate-cut expectations were scaled back as markets reassessed the likelihood of aggressive easing amid firm growth.
However, that move reversed decisively once the inflation data was released. Relatively tame CPI figures later in the week provided relief, particularly at the front end. Headline and core inflation both came in slightly cooler or in line with expectations, triggering a renewed repricing toward easing. Treasuries rallied broadly, with the curve flattening rather than steepening. The policy-sensitive two-year yield fell 9 basis points to 3.41%, its lowest level since late November, while the 10-year declined roughly 15 basis points to 4.05%, the lowest since early December. The front-end move fully unwound the earlier post-jobs spike, signalling that markets regained confidence in the disinflation narrative.
By week’s end, money markets were assigning roughly 50% odds of a third rate cut by December, and pricing higher probabilities that the Federal Reserve may cut rates more than twice this year. Wagers that slowing inflation will permit multiple cuts helped drive Treasury yields to their lowest levels of the year, contributing to what was described as the largest weekly gain in months.
Still, the tone of the broader discussion urges caution. January is widely regarded as one of the most seasonally distorted months for both inflation and employment data. The argument against aggressive repricing rests on the premise that a single CPI print, especially a January print, does not establish a durable trend. The base-case view remains closer to a single-cut glide path than to a multi-cut glide path.
The curve has been steepening, described as the path of least resistance. The front-end rally has outpaced longer maturities, reflecting shifting expectations around the timing and sequencing of cuts rather than a structural downgrade to long-term growth or inflation expectations. The 2-year continues to function as the cleanest proxy for Fed pricing.
Beyond the front end, Treasuries traded as a clean safe haven during episodes of uncertainty. The 30-year auction was described as exceptionally strong, with end-user demand reinforcing the bid for duration. This strength is notable given ongoing concerns about the fiscal deficit. Despite periodic headlines suggesting a retrenchment by official-sector buyers, the data do not indicate a meaningful structural shift away from US duration. Large global current account surpluses, particularly in East Asia, continue to require recycling into high-quality sovereign assets.
The dollar has softened alongside this repricing, though the move appears to reflect incremental diversification rather than capital flight. There is no evidence of broad US outflows. Instead, marginal capital appears to be rotating into Europe, Japan, and emerging markets, a “buy more world” dynamic rather than “sell America.” While rate differentials would typically argue for a firmer dollar, positioning and geopolitical considerations have temporarily weighed on the currency.
Two competing policy narratives remain in play. One view argues that policy may be tighter than necessary given supply-side improvements and residual slack. The other stresses that inflation remains above target, that growth is solid, and that January distortions complicate interpretation, thereby limiting the urgency of cuts. Markets have leaned toward the former; policymakers appear more cautious.
The net result is a rates market that is increasingly confident about easing but still highly sensitive to incremental data surprises. The front end reflects conviction. The long end reflects structural demand. And the curve reflects uncertainty about how quickly policy can normalise.
Credit
Credit markets remain liquid and well bid, but internal leadership has shifted, and the change is centred on technology.
Cross-currency issuance stayed exceptionally strong. A renewed wave of reverse Yankee supply hit Europe as U.S. issuers diversified funding bases and avoided overloading domestic curves. Alphabet led with a £5.5bn sterling deal alongside a record Swiss-franc print, followed by Goldman Sachs, JPMorgan and Ford. Goldman placed roughly €7bn, one of the largest euro-denominated financial deals on record. While USD yields around 4.8% versus roughly 3% in euros appear attractive, the primary motivation is access to broader liquidity pools amid a structurally heavy supply year.
U.S. investment-grade issuance totalled approximately $40bn. Alphabet accounted for $20bn via a seven-part deal that drew more than $100bn in orders, pricing from +60bp in 3 years to +90bp in 10 years. The bonds traded flat to 4–7bp tighter in secondary, with the strongest performance in shorter maturities. Oracle’s $25bn eight-part deal attracted a record $129bn of demand, yet its longer-dated bonds widened more than 18bp in secondary trading.
This front-end versus long-end bifurcation is the key dynamic. Investors remain comfortable with short-duration, high-quality tech exposure, but long-dated tech supply is being repriced. Roughly 40% of recent Alphabet and Oracle issuance came beyond 20 years, including Alphabet’s 100-year sterling tranche. The curve is not steep enough to compensate for added spread duration and uncertainty around long-term leverage and capital intensity. Duration repricing, not funding stress, is driving divergence.
Technology is now the worst-performing segment of the IG market year-to-date. The Bloomberg US Corporate IG index has tightened 3bp YTD and sits near 75bp OAS with a 4.83% yield-to-worst, but tech spreads have widened relative to the broader market. Cyclical sectors such as energy, materials, autos, and industrials have tightened and are currently offsetting the weakness in technology. If technology widening continues, it will begin to act as a direct drag on headline spreads rather than being absorbed internally.
The driver is structural. AI capex represents a multi-year financing cycle, not a cyclical issuance spike. Hyperscalers are funding data centres, chips and power infrastructure across currencies and maturities. Alphabet alone raised nearly $32bn across markets within 24 hours. This financing extends beyond IG corporates into leveraged loans, HY, asset-based finance and private credit. The market is effectively re-underwriting 10-20 years of cash flows across sectors simultaneously.
High yield continues to absorb AI-linked infrastructure risk. Spreads moved toward 275bp with yield-to-worst near 6.6%, and Track Capital’s $3.8bn junk bond tied to a Nevada data centre expected to be leased to Nvidia drew roughly $14bn in orders and priced tighter after upsizing. More than $80bn of HY bonds trade above call prices, supporting ongoing refinancing.
Dispersion remains elevated. Index spreads are contained, but beneath the surface, stronger credits trade above par, while a growing tail across software and related sectors faces pressure. Technology is no longer the stabilising core of IG credit. It has become the sector in which duration, leverage trajectory, and AI-driven capital intensity are being repriced first, and this process is increasingly relevant to index direction into 2026.
Equities
US equities ended the week lower, extending the recent consolidation across major benchmarks. The Dow declined 1.23%, the S&P 500 fell 1.39% for a second consecutive weekly loss, the Nasdaq dropped 2.10%, marking its fifth straight weekly decline, and the Russell 2000 slipped 0.89%, now negative in three of the past four weeks. Beneath the surface, however, the dominant theme was dispersion and rotation. The equal-weight S&P 500 outperformed the cap-weighted index by nearly 170 bp during the week and briefly set a fresh record high, reinforcing the ongoing broadening trade away from mega-cap concentration and toward cyclicals. Utilities (+7.13%), real estate (+3.87%), materials (+3.66%), and energy (+1.68%) led sector performance, while financials (−4.83%), communication services (−3.53%), consumer discretionary (−2.10%), and technology (−1.98%) lagged.
Earnings season continues to provide an important counterweight to index-level weakness. According to FactSet, the blended Q4 S&P 500 EPS growth rate stands at 13.2%, well above the 8.3% expected at the end of the quarter, with blended revenue growth at 9.0%. Among the 74% of companies that have reported, 74% have exceeded EPS estimates, below the 79% one-year and 78% five-year averages, while 73% have exceeded revenue expectations, above the 71% one-year and 70% five-year averages. In aggregate, earnings are 7.2% above expectations, slightly below the 7.4% one-year and 7.7% five-year positive surprise rates, while revenues are 1.6% above expectations, ahead of the 1.3% one-year average but below the 2.0% five-year norm. Within technology, divergence remains pronounced: mega-cap weakness pressured headline indices, yet AI infrastructure and semiconductor names have been comparatively resilient amid sustained AI-related capex demand, even as memory dynamics create both revenue tailwinds and margin headwinds across hardware. Software has staged only a modest rebound from its recent drawdown amid ongoing AI-displacement concerns, leaving the broader picture one of rotation and capital reallocation rather than deteriorating fundamentals.
The internal rotation within technology is no longer subtle, but decisive. Semiconductor equities have outperformed the broader tech-heavy Nasdaq 100 by roughly 50 percentage points since the start of 2025, with hardware proxies rising close to 68 per cent versus approximately 18 per cent for the index. That scale of relative performance reflects a market conviction that incremental AI spending is disproportionately allocated to compute, accelerators, memory, and data centre infrastructure rather than being evenly distributed across the technology ecosystem. Capital is rewarded by scarcity in physical AI capacity.
The divergence becomes even more striking when hardware is compared directly to software. Semiconductor indices are up roughly 58%, while software, represented by IGV, is down more than 22%, producing an extreame 80% performance gap. Investors are pricing the AI stack vertically. Layers closest to usage intensity and infrastructure scaling are enjoying high multiples and earnings growth. Application software, particularly horizontal workflow tools, faces a complex set of risks. AI enhances products but also lowers development barriers and compresses differentiation across most categories.
Implication is structural rather than tactical. If AI-driven productivity gains are real and increasingly embedded in corporate workflows, then the economic rent accrues first to the providers of compute and foundational infrastructure. Software must prove that AI integration expands its total addressable market and pricing power rather than eroding its moat. The widening performance gap is effectively the market’s judgment on where durable free cash flow growth is most visible today. Until application vendors demonstrate sustained margin expansion from AI rather than feature commoditization, hardware remains the more direct and defensible monetisation channel of the AI cycle.
You can connect it by framing the software sell-off as the equity market expression of the same structural shift described earlier in the productivity and semiconductor sections. The productivity acceleration and hardware outperformance are the economic cause; the software derating is the valuation consequence.
The extreme divergence across technologies and the emerging evidence of AI-driven productivity gains provide the economic backdrop for the sharp repricing of software. If AI is genuinely compressing task time, reducing entry-level labour demand by roughly 16% in exposed sectors, and enabling end-to-end workflow automation, then the layers capturing incremental economic rent will not be those built primarily on interface lock-in. The equity market is internalising that shift. Nearly $1 trillion has been erased from software and services equities in a matter of weeks. The S&P 500 Software & Services Index is down around 20% YTD. FactSet has fallen from a $20 billion peak to below $8 billion; S&P Global declined by roughly 30% within weeks; and Thomson Reuters has lost close to half its market capitalisation over the past year. This is not random volatility. It is the market adjusting terminal value assumptions in light of a new competitive reality.
The catalyst was the release of industry-specific plugins for Claude’s Cowork, an autonomous agent capable of handling complex research, analysis and document workflows. That development crystallised what had been theoretical. If natural language becomes the dominant interface and domain logic can be encoded directly into model-driven agents, then the moats that justified 15 to 20 times revenue multiples weaken materially. Importantly, enterprise revenue does not vanish overnight, given multi-year contracts and slow procurement cycles. The adjustment occurs first in the valuation, not in the income statement. The selloff, therefore, represents multiple compressions driven by perceived erosion of pricing power and switching costs, rather than an immediate collapse in demand. In that context, the hardware outperformance and the software derating are two sides of the same structural AI transition.