Stimulates hair follicles in a dish and reaches them from a shampoo, but proof that caffeine products slow hair loss in people is thin. Eye-cream claims are weaker still.
In hair products, caffeine is meant to counter the hormone-driven shrinking of follicles in pattern hair loss. It does reach follicles quickly from a shampoo, and it stimulates growth in follicles kept alive in a dish.
In eye creams and serums, caffeine is marketed to reduce puffiness and dark circles by narrowing blood vessels. That is plausible, but there is little trial evidence and any effect is short-lived.
Lab studies show caffeine stimulates human hair follicles in culture, and a pharmacokinetic study shows it enters follicles within 2 minutes of shampooing. The only sizeable clinical trial (210 men) was open-label, had no placebo arm, and found a 0.2% leave-on caffeine liquid 'non-inferior' to 5% minoxidil on a growing-hair ratio, not hair count. No placebo-controlled trial shows caffeine shampoo reduces hair loss.
In hair follicles taken from men with pattern hair loss and grown in a dish, caffeine at 0.001-0.005% counteracted testosterone's growth suppression and stimulated hair shaft growth on its own.
Caffeine from a shampoo left on for 2 minutes reached the blood, entering faster via hair follicles than through the skin surface. Shows delivery is possible, not that it grows hair.
210 men used a 0.2% caffeine leave-on liquid or 5% minoxidil for 6 months; the share of growing-phase hairs rose by 10.6% and 11.7% respectively, and caffeine was judged non-inferior. Open-label, no placebo, and a trichogram rather than hair-count endpoint.
Reviews caffeine's skin penetration and cosmetic uses, noting topical products usually contain about 3%: fat breakdown in anti-cellulite products, antioxidant and UV-protective effects, increased microcirculation and hair-growth stimulation, largely from lab data.